[V0 Deprecation] Remove V0 sampling metadata (#25345)
Signed-off-by: Woosuk Kwon <woosuk@thinkingmachines.ai>
This commit is contained in:
@@ -25,7 +25,6 @@ from vllm.model_executor.layers.quantization.base_config import (
|
||||
QuantizationConfig)
|
||||
from vllm.model_executor.layers.vocab_parallel_embedding import (
|
||||
DEFAULT_VOCAB_PADDING_SIZE, ParallelLMHead, VocabParallelEmbedding)
|
||||
from vllm.model_executor.sampling_metadata import SamplingMetadata
|
||||
from vllm.sequence import IntermediateTensors
|
||||
|
||||
from .granitemoe import GraniteMoeAttention, GraniteMoeModel, GraniteMoeMoE
|
||||
@@ -311,11 +310,9 @@ class GraniteMoeSharedForCausalLM(nn.Module, SupportsLoRA, SupportsPP):
|
||||
inputs_embeds)
|
||||
return hidden_states
|
||||
|
||||
def compute_logits(
|
||||
self, hidden_states: torch.Tensor,
|
||||
sampling_metadata: SamplingMetadata) -> Optional[torch.Tensor]:
|
||||
logits = self.logits_processor(self.lm_head, hidden_states,
|
||||
sampling_metadata)
|
||||
def compute_logits(self,
|
||||
hidden_states: torch.Tensor) -> Optional[torch.Tensor]:
|
||||
logits = self.logits_processor(self.lm_head, hidden_states)
|
||||
return logits
|
||||
|
||||
def make_empty_intermediate_tensors(
|
||||
|
||||
Reference in New Issue
Block a user